Importing Libraries

# Data pre-processing
import pandas as pd
import numpy as np
import missingno as msno
import warnings
warnings.filterwarnings("ignore")

#Data visualization
import matplotlib.pyplot as plt
import seaborn as sns
import plotly as pl
import plotly.express as px
import plotly.graph_objects as go
import plotly.figure_factory as ff
from plotly.subplots import make_subplots
%matplotlib inline

# Data Clustering
import sklearn
from sklearn import neighbors
from sklearn.preprocessing import StandardScaler
from sklearn.cluster import KMeans

Data Representation

#Data importing
countryvaccination = pd.read_csv("2021-05-19_country_vaccinations.csv")
cv = countryvaccination.copy()
countryvaccinationbyMan = pd.read_csv("2021-05-19_country_vaccinations_by_manufacturer.csv")
manuVac = countryvaccinationbyMan.copy()

#Converting to date format
vacc = countryvaccination.copy()
vacc['date'] = pd.to_datetime(vacc['date'], format='%Y-%m-%d')
vacc=vacc.set_index('date')

# Looking for the missing values
msno.bar(manuVac)  

plt.show()

 #Visualizing the missing data
msno.bar(vacc) 

plt.show()

Usecase 1: Which countries have the best vaccination progress?

#Taking the maximim of vaccination counts
countryvaccination1 = countryvaccination.groupby(["country",'iso_code','vaccines'])['total_vaccinations',
                     'people_vaccinated','people_fully_vaccinated',
                    'daily_vaccinations','total_vaccinations_per_hundred',
                    'people_vaccinated_per_hundred',"people_fully_vaccinated_per_hundred"
,'daily_vaccinations_per_million'].max().reset_index()

countryvaccinationbyMan1 = countryvaccinationbyMan.groupby(["location",'vaccine'])['total_vaccinations'].max().reset_index()

countryvaccination1

	country	iso_code	vaccines	total_vaccinations	people_vaccinated	people_fully_vaccinated	daily_vaccinations	total_vaccinations_per_hundred	people_vaccinated_per_hundred	people_fully_vaccinated_per_hundred	daily_vaccinations_per_million
0	Afghanistan	AFG	Oxford/AstraZeneca, Pfizer/BioNTech, Sinopharm...	504502.0	448878.0	55624.0	13921.0	1.30	1.15	0.14	358.0
1	Albania	ALB	Oxford/AstraZeneca, Pfizer/BioNTech, Sinovac, ...	632676.0	444755.0	187921.0	17565.0	21.98	15.45	6.53	6104.0
2	Algeria	DZA	Sputnik V	75000.0	NaN	NaN	3748.0	0.17	NaN	NaN	85.0
3	Andorra	AND	Oxford/AstraZeneca, Pfizer/BioNTech	31633.0	26931.0	4702.0	1182.0	40.94	34.86	6.09	15298.0
4	Angola	AGO	Oxford/AstraZeneca	626572.0	586377.0	40195.0	18751.0	1.91	1.78	0.12	571.0
...	...	...	...	...	...	...	...	...	...	...	...
206	Wales	OWID_WLS	Moderna, Oxford/AstraZeneca, Pfizer/BioNTech	2853576.0	1975820.0	877756.0	33151.0	90.51	62.67	27.84	10515.0
207	Wallis and Futuna	WLF	Oxford/AstraZeneca	5774.0	4286.0	1488.0	272.0	51.34	38.11	13.23	24186.0
208	Yemen	YEM	Oxford/AstraZeneca	18555.0	18555.0	NaN	NaN	0.06	0.06	NaN	NaN
209	Zambia	ZMB	Oxford/AstraZeneca	90916.0	90916.0	NaN	5680.0	0.49	0.49	NaN	309.0
210	Zimbabwe	ZWE	Sinopharm/Beijing	730365.0	549797.0	180568.0	22863.0	4.91	3.70	1.21	1538.0

211 rows × 11 columns

# Visualizing the total vaccinations throgh Choropleth

# We are plotting total vaccinations throughout the world and color bar indicates the count of total vaccinations.
# From this plot we can observe that China,UnitedStates and India has the total number of vaccinations count.
#But to take population count also into account to decide on best vaccination progress. 
from plotly.offline import download_plotlyjs, init_notebook_mode, plot, iplot
trace = go.Choropleth(
            locations = countryvaccination1['country'],
            locationmode='country names',
            z = countryvaccination1['total_vaccinations'],
            text = countryvaccination1['country'],
            autocolorscale =False,
            reversescale = True,
            colorscale = 'viridis',
            marker = dict(
                line = dict(
                    color = 'rgb(0,0,0)',
                    width = 0.5)
            ),
            colorbar = dict(
                title = 'Total vaccinations',
                tickprefix = '')
        )

data = [trace]
layout = go.Layout(
    title = 'Total vaccinations for country',
    geo = dict(
        showframe = True,
        showlakes = False,
        showcoastlines = True,
        projection = dict(
            type = 'natural earth'
        )
    )
)
fig = dict( data=data, layout=layout )
iplot(fig)

# Tree map of vaccinations used by respective country and its total vaccination count.From this plot,
#we can see the type of vaccinations used by top vaccinating countries.

fig = px.treemap(countryvaccination1,values = 'total_vaccinations',
                 path = ['vaccines','country'],
                 title="Total Vaccinations for a country")
fig.show()

#Top 10 countries which has the best vaccination rate from december 2020 to may 2021. 
#The countries which are seen through legends of a plot.

fig = px.area(countryvaccination[countryvaccination.country.isin(countryvaccination[countryvaccination['date'] == 
countryvaccination['date'].max()].sort_values("total_vaccinations", ascending=False).iloc[:10].country)].sort_values("total_vaccinations", ascending=False),
x="date", y="total_vaccinations", color="country", template="plotly_dark")

fig.update_traces(line={"width":1.25})
fig.update_layout(title = f"Top 10 Countries with Most Vaccinated on {countryvaccination['date'].max()}",
xaxis={"title": "date"}, yaxis={"title": "total vaccinations" })

# Bar graph of total vaccinations per 100 for top 20 countries. From this plot we get the fastest vaccinating 
# countries but with less population like Gibratar, Seychelies and Falkland Islands have given vaccinations to almost everyone.

x= countryvaccination1.groupby("country")["total_vaccinations_per_hundred"].max().sort_values(ascending= False).head(20)
plt.figure(figsize= (15,10))
ax= sns.barplot(x.values,x.index)
ax.set_xlabel("total vaccinations")
ax.set_ylabel("Country")
plt.show()

#Importing the world population dataset by worldometer
pop = pd.read_csv("population_by_country_2020.csv")

pop['country'] = pop['Country (or dependency)']
pop1 = pop[['country','Population (2020)']]

pop1.reset_index(drop=True)

	country	Population (2020)
0	China	1440297825
1	India	1382345085
2	United States	331341050
3	Indonesia	274021604
4	Pakistan	221612785
...	...	...
230	Montserrat	4993
231	Falkland Islands	3497
232	Niue	1628
233	Tokelau	1360
234	Holy See	801

235 rows × 2 columns

# merging the population count to the vaccinations dataframe
cc= countryvaccination1.copy()
countryvaccination1 = countryvaccination1.merge(pop1,on='country')

#Calculating the vaccination rate of all the countries
countryvaccination1['percentage_fully_vaccinated'] = countryvaccination1['people_fully_vaccinated'].fillna(0) *100 / countryvaccination1['Population (2020)']
countryvaccination1['percentage_vaccinated'] = countryvaccination1['total_vaccinations'].fillna(0)/ countryvaccination1['Population (2020)']*100

#Plotting the vaccination rate of all countries and the color bar indicates the percentage of vaccinated. 
#From this plot we can tell that europe has the highest people vaccinated according to their population rate.
fig = px.scatter_geo(countryvaccination1, locations="country", color="percentage_vaccinated",
                     locationmode='country names',
                     hover_name="country", size="percentage_vaccinated",
                     projection="natural earth")
fig.update_layout(title='The number of vaccinated people per population', title_x=0.45)
fig.show()

# bar plot indicates Daily vaccination count for top 10 countries. 
#From this plot we can observe that China, USA ad India are the ones with high number of daily vaccination rate.

daily= countryvaccination.groupby("country").daily_vaccinations.mean().sort_values(ascending= False).head(10)
plt.figure(figsize= (15,10))
ax= sns.barplot(daily.values,daily.index)
ax.set_title("Daily Vaccinations Count")
ax.set_xlabel("Daily Vaccinations(avg)")
ax.set_ylabel("Country")
plt.show()

From the above plots, We can see the countries having highest vaccination rate for 100 and total. We oberved that top cuntries having good progress are China,USA and India. By taking the vaccination rate and population count, we can tell that United States has the best Vaccination progress across the world.

# Creating the dataframes for top vaccinating countries.
China = countryvaccination[countryvaccination['country']=='China']
UnitedStates = countryvaccination[countryvaccination['country']=='United States']
India = countryvaccination[countryvaccination['country']=='India']
Germany = countryvaccination[countryvaccination['country']=='Germany']
UK = countryvaccination[countryvaccination['country']=='United Kingdom']

Time Series Forecast

China.to_csv("China.csv")

China1 = pd.read_csv('China.csv',index_col='date',parse_dates=True)
China1['total_vaccinations']= China1['total_vaccinations'].fillna(method='ffill').fillna(method='bfill')
China1['daily_vaccinations'] = China1['daily_vaccinations'].fillna(0)

China2 = China1[['daily_vaccinations']]
China3 = China1[['total_vaccinations']]

China1['daily_vaccinations'].plot(figsize=(12,6))

<AxesSubplot:xlabel='date'>

China2 = China2.dropna()

#Visuaizing the total vaccination, We can observe that there is no trend or seasonality
#in this plot to do the prediction
plt.figure(figsize=(10,6))
plt.plot(China2.diff())

[<matplotlib.lines.Line2D at 0x1e359c31d60>]

##Visuaizing the total vaccination difference, We can observe that there is no trend or seasonality
#in this plot to do the prediction. Hence we are ignoring this country for forecasting.
plt.figure(figsize=(10,6))
plt.plot(China3.diff())

[<matplotlib.lines.Line2D at 0x1e359c9cdc0>]

USA Daily and Total Vaccination timeseries forecast

#Importing UnitedStates dataframe and pre-processing it.
UnitedStates.to_csv("UnitedStates.csv")
UnitedStates1 = pd.read_csv('UnitedStates.csv',index_col='date',parse_dates=True)
UnitedStates1['daily_vaccinations'] = UnitedStates1.daily_vaccinations.fillna(0)
UnitedStates1['total_vaccinations'] = UnitedStates1.total_vaccinations.fillna(method='ffill')

UnitedStates1.info()

<class 'pandas.core.frame.DataFrame'>
DatetimeIndex: 145 entries, 2020-12-20 to 2021-05-13
Data columns (total 15 columns):
 #   Column                               Non-Null Count  Dtype  
---  ------                               --------------  -----  
 0   Unnamed: 0                           145 non-null    int64  
 1   country                              145 non-null    object 
 2   iso_code                             145 non-null    object 
 3   total_vaccinations                   145 non-null    float64
 4   people_vaccinated                    130 non-null    float64
 5   people_fully_vaccinated              116 non-null    float64
 6   daily_vaccinations_raw               121 non-null    float64
 7   daily_vaccinations                   145 non-null    float64
 8   total_vaccinations_per_hundred       131 non-null    float64
 9   people_vaccinated_per_hundred        130 non-null    float64
 10  people_fully_vaccinated_per_hundred  116 non-null    float64
 11  daily_vaccinations_per_million       144 non-null    float64
 12  vaccines                             145 non-null    object 
 13  source_name                          145 non-null    object 
 14  source_website                       145 non-null    object 
dtypes: float64(9), int64(1), object(5)
memory usage: 18.1+ KB

UnitedStates2 = UnitedStates1[['total_vaccinations']]

#Plotting the total vaccination difference. From thsi plot we can see the seasoanl trend to analyze and do the prediction.
plt.figure(figsize=(10,6))
plt.plot(UnitedStates2.diff())

[<matplotlib.lines.Line2D at 0x1e35a48bb50>]

UnitedStates2=UnitedStates2.diff()
UnitedStates2=UnitedStates2.dropna()

#Splitting the dataset for training and testing.
start_date='2021-04-12'

train=UnitedStates2.loc[UnitedStates2.index < pd.to_datetime(start_date)]
test=UnitedStates2.loc[UnitedStates2.index >= pd.to_datetime(start_date)]

# Training the SARIMAX model for the time-series forecast analysis. Using this model,we can do the future prediction on total 
# Vaccination count after a month of a may for 30 days.
from statsmodels.tsa.statespace.sarimax import SARIMAX
model=SARIMAX(UnitedStates2, order=(5, 1, 4))
results=model.fit(disp=True)

c:\python\lib\site-packages\statsmodels\tsa\base\tsa_model.py:159: ValueWarning:

No frequency information was provided, so inferred frequency D will be used.

c:\python\lib\site-packages\statsmodels\tsa\base\tsa_model.py:159: ValueWarning:

No frequency information was provided, so inferred frequency D will be used.

c:\python\lib\site-packages\statsmodels\base\model.py:567: ConvergenceWarning:

Maximum Likelihood optimization failed to converge. Check mle_retvals

# The predicted plot for the month of may. We can oberve the orange line representing the prediction and the 
#blue as the true values. 

sarimax_prediction=results.predict(
start=start_date, end='2021-05-13', dynamic=False)
plt.figure(figsize=(10,5))
l1,=plt.plot(UnitedStates2, label='Observation')
l2,=plt.plot(sarimax_prediction, label='ARIMA')
plt.legend(handles=[l1,l2])
plt.ticklabel_format(style="plain", axis = 'y')
plt.savefig('SARIMAX prediction', bbox_inches='tight', transparent=False)

# Running auto arima for to find the season order.
import pmdarima as pm
stepwise = pm.auto_arima(UnitedStates2,trace=True,suppress_warnings=True)
stepwise.summary()

Performing stepwise search to minimize aic
 ARIMA(2,1,2)(0,0,0)[0] intercept   : AIC=4258.520, Time=0.56 sec
 ARIMA(0,1,0)(0,0,0)[0] intercept   : AIC=4302.524, Time=0.01 sec
 ARIMA(1,1,0)(0,0,0)[0] intercept   : AIC=4293.495, Time=0.02 sec
 ARIMA(0,1,1)(0,0,0)[0] intercept   : AIC=4269.341, Time=0.02 sec
 ARIMA(0,1,0)(0,0,0)[0]             : AIC=4300.564, Time=0.02 sec
 ARIMA(1,1,2)(0,0,0)[0] intercept   : AIC=4267.529, Time=0.06 sec
 ARIMA(2,1,1)(0,0,0)[0] intercept   : AIC=4262.775, Time=0.09 sec
 ARIMA(3,1,2)(0,0,0)[0] intercept   : AIC=4229.788, Time=0.20 sec
 ARIMA(3,1,1)(0,0,0)[0] intercept   : AIC=4251.779, Time=0.10 sec
 ARIMA(4,1,2)(0,0,0)[0] intercept   : AIC=4217.919, Time=0.32 sec
 ARIMA(4,1,1)(0,0,0)[0] intercept   : AIC=4236.414, Time=0.10 sec
 ARIMA(5,1,2)(0,0,0)[0] intercept   : AIC=4209.518, Time=0.70 sec
 ARIMA(5,1,1)(0,0,0)[0] intercept   : AIC=4220.284, Time=0.19 sec
 ARIMA(5,1,3)(0,0,0)[0] intercept   : AIC=4208.658, Time=0.78 sec
 ARIMA(4,1,3)(0,0,0)[0] intercept   : AIC=4215.307, Time=1.06 sec
 ARIMA(5,1,4)(0,0,0)[0] intercept   : AIC=4208.482, Time=1.74 sec
 ARIMA(4,1,4)(0,0,0)[0] intercept   : AIC=4213.159, Time=0.81 sec
 ARIMA(5,1,5)(0,0,0)[0] intercept   : AIC=4212.247, Time=0.77 sec
 ARIMA(4,1,5)(0,0,0)[0] intercept   : AIC=4216.972, Time=0.89 sec
 ARIMA(5,1,4)(0,0,0)[0]             : AIC=4214.991, Time=1.52 sec

Best model:  ARIMA(5,1,4)(0,0,0)[0] intercept
Total fit time: 9.935 seconds

SARIMAX Results

Dep. Variable:	y	No. Observations:	144
Model:	SARIMAX(5, 1, 4)	Log Likelihood	-2093.241
Date:	Tue, 01 Jun 2021	AIC	4208.482
Time:	02:27:51	BIC	4241.073
Sample:	0	HQIC	4221.725
	- 144
Covariance Type:	opg

	coef	std err	z	P>|z|	[0.025	0.975]
intercept	5.345e+04	2.82e+04	1.897	0.058	-1783.588	1.09e+05
ar.L1	0.2004	0.510	0.393	0.694	-0.798	1.199
ar.L2	-0.2972	0.473	-0.629	0.530	-1.224	0.629
ar.L3	-0.5558	0.356	-1.560	0.119	-1.254	0.142
ar.L4	-0.2428	0.154	-1.575	0.115	-0.545	0.059
ar.L5	-0.2708	0.138	-1.962	0.050	-0.541	-0.000
ma.L1	-1.0979	0.517	-2.125	0.034	-2.110	-0.085
ma.L2	0.4795	0.923	0.520	0.603	-1.329	2.288
ma.L3	0.3094	0.822	0.376	0.707	-1.302	1.921
ma.L4	-0.2551	0.322	-0.793	0.428	-0.886	0.376
sigma2	3.097e+11	0.005	6.82e+13	0.000	3.1e+11	3.1e+11

Ljung-Box (Q):	22.81	Jarque-Bera (JB):	136.99
Prob(Q):	0.99	Prob(JB):	0.00
Heteroskedasticity (H):	0.42	Skew:	0.99
Prob(H) (two-sided):	0.00	Kurtosis:	7.36

Warnings:
[1] Covariance matrix calculated using the outer product of gradients (complex-step).
[2] Covariance matrix is singular or near-singular, with condition number 1.62e+29. Standard errors may be unstable.

# Performing Adfuller test to check for the stationary.

from statsmodels.tsa.stattools import adfuller
def ad_test(dataset):
    
    dftest=adfuller(dataset,autolag='AIC')
    print("1. ADF: ", dftest[0])
    print("2. P-Value: ", dftest[1])
    print("3. Num of Lags: ", dftest[2])
    print("4. Number of observations used for ADF Regression and critival values calculation: ", dftest[2])
    print("5. Critical Values: ", dftest[4])
    for key, val in dftest[4].items():
        print("\t",key,":",val)
ad_test(UnitedStates2)

1. ADF:  -1.9020246173694153
2. P-Value:  0.33110926706826116
3. Num of Lags:  9
4. Number of observations used for ADF Regression and critival values calculation:  9
5. Critical Values:  {'1%': -3.480118600110386, '5%': -2.8833618426136196, '10%': -2.578407034974382}
	 1% : -3.480118600110386
	 5% : -2.8833618426136196
	 10% : -2.578407034974382

UnitedStates2.tail(1)

	total_vaccinations
date
2021-05-13	1915642.0

# Fitting the model to predict the future of total vaccination count from may 14th to june 13th.

model2 = SARIMAX(UnitedStates2,order=(5,1,4))
model2= model2.fit(disp=True)
#model2.summary()
index_feature_dates=pd.date_range(start='2021-05-14', end='2021-06-13')
pred = model2.predict(start=len(UnitedStates2),end=len(UnitedStates2)+30).rename('total_vaccinations')
pred.index= index_feature_dates
print(pred)

c:\python\lib\site-packages\statsmodels\tsa\base\tsa_model.py:159: ValueWarning:

No frequency information was provided, so inferred frequency D will be used.

c:\python\lib\site-packages\statsmodels\tsa\base\tsa_model.py:159: ValueWarning:

No frequency information was provided, so inferred frequency D will be used.

2021-05-14    2.560729e+06
2021-05-15    2.687125e+06
2021-05-16    2.391035e+06
2021-05-17    1.882785e+06
2021-05-18    1.593458e+06
2021-05-19    1.729720e+06
2021-05-20    2.182042e+06
2021-05-21    2.599084e+06
2021-05-22    2.666382e+06
2021-05-23    2.334943e+06
2021-05-24    1.852494e+06
2021-05-25    1.573583e+06
2021-05-26    1.697578e+06
2021-05-27    2.126618e+06
2021-05-28    2.544150e+06
2021-05-29    2.650234e+06
2021-05-30    2.378217e+06
2021-05-31    1.936060e+06
2021-06-01    1.647689e+06
2021-06-02    1.715392e+06
2021-06-03    2.077716e+06
2021-06-04    2.462621e+06
2021-06-05    2.591133e+06
2021-06-06    2.379247e+06
2021-06-07    1.992097e+06
2021-06-08    1.716060e+06
2021-06-09    1.747395e+06
2021-06-10    2.053168e+06
2021-06-11    2.402725e+06
2021-06-12    2.541308e+06
2021-06-13    2.375372e+06
Freq: D, Name: total_vaccinations, dtype: float64

c:\python\lib\site-packages\statsmodels\base\model.py:567: ConvergenceWarning:

Maximum Likelihood optimization failed to converge. Check mle_retvals

# This graph represents the total vaccinatios difference graph after concatenating the predicted values to the dataframe

preditions = pd.DataFrame(pred)
preditions
frames = [UnitedStates2, preditions]
  
UnitedStates2 = pd.concat(frames)
UnitedStates2.plot()

<AxesSubplot:>

sum1 = pred.sum()
sum1

67088160.22308184

UnitedStates3 = UnitedStates1.groupby(["country",'iso_code','vaccines'])['total_vaccinations',
                     'people_vaccinated','people_fully_vaccinated',
                    'daily_vaccinations','total_vaccinations_per_hundred',
                    'people_vaccinated_per_hundred',"people_fully_vaccinated_per_hundred"
,'daily_vaccinations_per_million'].max().reset_index()

UnitedStates3

	country	iso_code	vaccines	total_vaccinations	people_vaccinated	people_fully_vaccinated	daily_vaccinations	total_vaccinations_per_hundred	people_vaccinated_per_hundred	people_fully_vaccinated_per_hundred	daily_vaccinations_per_million
0	United States	USA	Johnson&Johnson, Moderna, Pfizer/BioNTech	266596486.0	154624231.0	118987308.0	3384387.0	79.71	46.23	35.58	10120.0

sum2 = sum1 + UnitedStates3['total_vaccinations'].astype(int)
sum2

0    3.336846e+08
Name: total_vaccinations, dtype: float64

UnitedStates3['sum_daily_vaccinations_30days'] = sum1
UnitedStates3['total_vaccinations_after30days'] = sum2
UnitedStates3['total_vaccinations_after30days'] = UnitedStates3['total_vaccinations_after30days'].astype(int)
UnitedStates3

	country	iso_code	vaccines	total_vaccinations	people_vaccinated	people_fully_vaccinated	daily_vaccinations	total_vaccinations_per_hundred	people_vaccinated_per_hundred	people_fully_vaccinated_per_hundred	daily_vaccinations_per_million	sum_daily_vaccinations_30days	total_vaccinations_after30days
0	United States	USA	Johnson&Johnson, Moderna, Pfizer/BioNTech	266596486.0	154624231.0	118987308.0	3384387.0	79.71	46.23	35.58	10120.0	6.708816e+07	333684646

UnitedStates3['Vaccination rate'] =(1 - (UnitedStates3['total_vaccinations'] / UnitedStates3['total_vaccinations_after30days'])) * 100

# From this plot we can see the total vaccinations count after a month of may in the orange bars and the blue bar represents 
# the total vaccination count before prediction
UnitedStates3.plot(x="vaccines", y=["total_vaccinations", "total_vaccinations_after30days"], kind="bar")
plt.ticklabel_format(style="plain", axis = 'y')
plt.xticks(rotation = 360)
plt.legend(loc='upper left')

<matplotlib.legend.Legend at 0x1e35f0c9910>

# This figure represents the total vaccination rate increate in % after a month. We can see that 20% increase in the rate of total vaccination
# after a month.
fig = px.bar(UnitedStates3, x='country', y=["total_vaccinations", "sum_daily_vaccinations_30days"],
             hover_data=['Vaccination rate'], color='vaccines',
             title='Vaccination rate increase pred', height=400)
fig.show()

United Kingdom timeseries prediction

UK.to_csv("UK.csv")
UK1 = pd.read_csv('UK.csv',index_col='date',parse_dates=True)
UK1['daily_vaccinations'] = UK1.daily_vaccinations.fillna(0)
UK1['total_vaccinations'] = UK1.total_vaccinations.fillna(method='ffill')

UK2 = UK1[['total_vaccinations']]
plt.figure(figsize=(10,6))
plt.plot(UK2.diff())

[<matplotlib.lines.Line2D at 0x1e346cd4fd0>]

UK2=UK2.diff()
UK2=UK2.fillna(0)

start_date='2021-04-12'

train2=UK2.loc[UK2.index < pd.to_datetime(start_date)]
test2=UK2.loc[UK2.index >= pd.to_datetime(start_date)]

from statsmodels.tsa.statespace.sarimax import SARIMAX
model4=SARIMAX(UK2, order=(1, 1, 1))
results2=model4.fit(disp=True)

c:\python\lib\site-packages\statsmodels\tsa\base\tsa_model.py:159: ValueWarning:

No frequency information was provided, so inferred frequency D will be used.

c:\python\lib\site-packages\statsmodels\tsa\base\tsa_model.py:159: ValueWarning:

No frequency information was provided, so inferred frequency D will be used.

sarimax_prediction2=results2.predict(
start=start_date, end='2021-05-13', dynamic=False)
plt.figure(figsize=(10,5))
l1,=plt.plot(UK2, label='Observation')
l2,=plt.plot(sarimax_prediction2, label='ARIMA')
plt.legend(handles=[l1,l2])
plt.ticklabel_format(style="plain", axis = 'y')
plt.savefig('SARIMAX prediction', bbox_inches='tight', transparent=False)

import pmdarima as pm
stepwise = pm.auto_arima(UK2,trace=True,suppress_warnings=True)
stepwise.summary()

Performing stepwise search to minimize aic
 ARIMA(2,1,2)(0,0,0)[0] intercept   : AIC=inf, Time=0.33 sec
 ARIMA(0,1,0)(0,0,0)[0] intercept   : AIC=3512.991, Time=0.02 sec
 ARIMA(1,1,0)(0,0,0)[0] intercept   : AIC=3504.356, Time=0.02 sec
 ARIMA(0,1,1)(0,0,0)[0] intercept   : AIC=3488.613, Time=0.02 sec
 ARIMA(0,1,0)(0,0,0)[0]             : AIC=3511.078, Time=0.00 sec
 ARIMA(1,1,1)(0,0,0)[0] intercept   : AIC=3484.388, Time=0.05 sec
 ARIMA(2,1,1)(0,0,0)[0] intercept   : AIC=3484.574, Time=0.23 sec
 ARIMA(1,1,2)(0,0,0)[0] intercept   : AIC=3530.320, Time=0.16 sec
 ARIMA(0,1,2)(0,0,0)[0] intercept   : AIC=3522.697, Time=0.11 sec
 ARIMA(2,1,0)(0,0,0)[0] intercept   : AIC=3502.052, Time=0.04 sec
 ARIMA(1,1,1)(0,0,0)[0]             : AIC=3478.238, Time=0.10 sec
 ARIMA(0,1,1)(0,0,0)[0]             : AIC=3486.621, Time=0.05 sec
 ARIMA(1,1,0)(0,0,0)[0]             : AIC=3502.488, Time=0.02 sec
 ARIMA(2,1,1)(0,0,0)[0]             : AIC=3479.727, Time=0.21 sec
 ARIMA(1,1,2)(0,0,0)[0]             : AIC=3515.292, Time=0.16 sec
 ARIMA(0,1,2)(0,0,0)[0]             : AIC=3509.306, Time=0.09 sec
 ARIMA(2,1,0)(0,0,0)[0]             : AIC=3500.222, Time=0.05 sec
 ARIMA(2,1,2)(0,0,0)[0]             : AIC=inf, Time=0.64 sec

Best model:  ARIMA(1,1,1)(0,0,0)[0]          
Total fit time: 2.282 seconds

SARIMAX Results

Dep. Variable:	y	No. Observations:	130
Model:	SARIMAX(1, 1, 1)	Log Likelihood	-1736.119
Date:	Tue, 01 Jun 2021	AIC	3478.238
Time:	02:27:57	BIC	3486.817
Sample:	0	HQIC	3481.724
	- 130
Covariance Type:	opg

	coef	std err	z	P>|z|	[0.025	0.975]
ar.L1	0.3577	0.072	4.958	0.000	0.216	0.499
ma.L1	-0.9218	0.055	-16.857	0.000	-1.029	-0.815
sigma2	3.368e+10	2.03e-12	1.66e+22	0.000	3.37e+10	3.37e+10

Ljung-Box (Q):	90.49	Jarque-Bera (JB):	2636.20
Prob(Q):	0.00	Prob(JB):	0.00
Heteroskedasticity (H):	0.38	Skew:	2.62
Prob(H) (two-sided):	0.00	Kurtosis:	24.52

Warnings:
[1] Covariance matrix calculated using the outer product of gradients (complex-step).
[2] Covariance matrix is singular or near-singular, with condition number 3.09e+37. Standard errors may be unstable.

Germany Timeseries Prediction

Germany.to_csv("Germany.csv")
Germany1 = pd.read_csv('Germany.csv',index_col='date',parse_dates=True)
Germany1['daily_vaccinations'] = Germany1.daily_vaccinations.fillna(0)
Germany1['total_vaccinations'] = Germany1.total_vaccinations.fillna(method='ffill')

Germany2 = Germany1[['total_vaccinations']]

plt.figure(figsize=(10,6))
plt.plot(Germany2.diff())

[<matplotlib.lines.Line2D at 0x1e3600be9d0>]

Germany2=Germany2.diff()
Germany2=Germany2.fillna(0)

start_date='2021-04-12'

train1=Germany2.loc[Germany2.index < pd.to_datetime(start_date)]
test1=Germany2.loc[Germany2.index >= pd.to_datetime(start_date)]

from statsmodels.tsa.statespace.sarimax import SARIMAX
model1=SARIMAX(Germany2, order=(4, 1, 3))
results1=model1.fit(disp=True)

c:\python\lib\site-packages\statsmodels\tsa\base\tsa_model.py:159: ValueWarning:

No frequency information was provided, so inferred frequency D will be used.

c:\python\lib\site-packages\statsmodels\tsa\base\tsa_model.py:159: ValueWarning:

No frequency information was provided, so inferred frequency D will be used.

c:\python\lib\site-packages\statsmodels\base\model.py:567: ConvergenceWarning:

Maximum Likelihood optimization failed to converge. Check mle_retvals

sarimax_prediction1=results1.predict(
start=start_date, end='2021-05-13', dynamic=False)
plt.figure(figsize=(10,5))
l1,=plt.plot(Germany2, label='Observation')
l2,=plt.plot(sarimax_prediction1, label='ARIMA')
plt.legend(handles=[l1,l2])
plt.ticklabel_format(style="plain", axis = 'y')
plt.savefig('SARIMAX prediction', bbox_inches='tight', transparent=False)

import pmdarima as pm
stepwise = pm.auto_arima(Germany2,trace=True,suppress_warnings=True)
stepwise.summary()

Performing stepwise search to minimize aic
 ARIMA(2,1,2)(0,0,0)[0] intercept   : AIC=3547.874, Time=0.18 sec
 ARIMA(0,1,0)(0,0,0)[0] intercept   : AIC=3655.540, Time=0.00 sec
 ARIMA(1,1,0)(0,0,0)[0] intercept   : AIC=3647.933, Time=0.02 sec
 ARIMA(0,1,1)(0,0,0)[0] intercept   : AIC=3641.569, Time=0.04 sec
 ARIMA(0,1,0)(0,0,0)[0]             : AIC=3653.599, Time=0.01 sec
 ARIMA(1,1,2)(0,0,0)[0] intercept   : AIC=3622.635, Time=0.05 sec
 ARIMA(2,1,1)(0,0,0)[0] intercept   : AIC=3548.791, Time=0.11 sec
 ARIMA(3,1,2)(0,0,0)[0] intercept   : AIC=3544.223, Time=0.18 sec
 ARIMA(3,1,1)(0,0,0)[0] intercept   : AIC=3535.895, Time=0.18 sec
 ARIMA(3,1,0)(0,0,0)[0] intercept   : AIC=3554.664, Time=0.10 sec
 ARIMA(4,1,1)(0,0,0)[0] intercept   : AIC=3527.844, Time=0.32 sec
 ARIMA(4,1,0)(0,0,0)[0] intercept   : AIC=inf, Time=0.14 sec
 ARIMA(5,1,1)(0,0,0)[0] intercept   : AIC=3529.759, Time=0.75 sec
 ARIMA(4,1,2)(0,0,0)[0] intercept   : AIC=3525.164, Time=0.40 sec
 ARIMA(5,1,2)(0,0,0)[0] intercept   : AIC=3526.371, Time=0.93 sec
 ARIMA(4,1,3)(0,0,0)[0] intercept   : AIC=3499.934, Time=1.11 sec
 ARIMA(3,1,3)(0,0,0)[0] intercept   : AIC=3539.343, Time=0.66 sec
 ARIMA(5,1,3)(0,0,0)[0] intercept   : AIC=3501.807, Time=0.67 sec
 ARIMA(4,1,4)(0,0,0)[0] intercept   : AIC=3505.564, Time=0.65 sec
 ARIMA(3,1,4)(0,0,0)[0] intercept   : AIC=3536.660, Time=0.37 sec
 ARIMA(5,1,4)(0,0,0)[0] intercept   : AIC=3506.528, Time=0.98 sec
 ARIMA(4,1,3)(0,0,0)[0]             : AIC=3502.516, Time=1.21 sec

Best model:  ARIMA(4,1,3)(0,0,0)[0] intercept
Total fit time: 9.057 seconds

SARIMAX Results

Dep. Variable:	y	No. Observations:	138
Model:	SARIMAX(4, 1, 3)	Log Likelihood	-1740.967
Date:	Tue, 01 Jun 2021	AIC	3499.934
Time:	02:28:08	BIC	3526.214
Sample:	0	HQIC	3510.613
	- 138
Covariance Type:	opg

	coef	std err	z	P>|z|	[0.025	0.975]
intercept	1.272e+04	9754.945	1.304	0.192	-6400.124	3.18e+04
ar.L1	-0.7155	0.066	-10.889	0.000	-0.844	-0.587
ar.L2	-0.0087	0.055	-0.158	0.874	-0.116	0.099
ar.L3	-0.6823	0.058	-11.688	0.000	-0.797	-0.568
ar.L4	-0.9627	0.044	-21.755	0.000	-1.049	-0.876
ma.L1	0.9367	0.141	6.655	0.000	0.661	1.213
ma.L2	-0.4434	0.149	-2.973	0.003	-0.736	-0.151
ma.L3	-0.6169	0.138	-4.477	0.000	-0.887	-0.347
sigma2	5.967e+09	0.035	1.72e+11	0.000	5.97e+09	5.97e+09

Ljung-Box (Q):	35.70	Jarque-Bera (JB):	2618.35
Prob(Q):	0.66	Prob(JB):	0.00
Heteroskedasticity (H):	22.72	Skew:	-2.41
Prob(H) (two-sided):	0.00	Kurtosis:	23.87

Warnings:
[1] Covariance matrix calculated using the outer product of gradients (complex-step).
[2] Covariance matrix is singular or near-singular, with condition number 3.56e+27. Standard errors may be unstable.

Usecase 2: Which countries are using which vaccines

# This plot represents country wise vaccination usage across the world. The colors on the map represents 
# the vaccines used for each country.
# For example, Germany uses the vaccines like jhonson & jhonson, Moderna, pfizer and Astrazeneca represented in light green 
# color.
# Similarly the countries like France and Spain use the same vaccinations.

fig = px.choropleth(countryvaccination1, locations="iso_code",
                    color="vaccines",
                    hover_name="country",
                    color_continuous_scale=px.colors.sequential.Plasma,
                   title= "Vaccines used by different countries")

fig.update_layout(showlegend=False)
fig.show()

# This shows the vaccines and it is used by countries count. We can oberve that Oxford/AstraZeneca is
# used by most of the countries.
vaccine = countryvaccination1["vaccines"].value_counts().reset_index()
vaccine.columns = ['popular Vaccines','Number of Countries']
vaccine.nlargest(10,'Number of Countries')

	popular Vaccines	Number of Countries
0	Oxford/AstraZeneca	52
1	Johnson&Johnson, Moderna, Oxford/AstraZeneca, ...	13
2	Moderna, Oxford/AstraZeneca, Pfizer/BioNTech	12
3	Oxford/AstraZeneca, Pfizer/BioNTech	11
4	Pfizer/BioNTech	10
5	Oxford/AstraZeneca, Sinopharm/Beijing	9
6	Oxford/AstraZeneca, Pfizer/BioNTech, Sinovac	8
7	Oxford/AstraZeneca, Pfizer/BioNTech, Sinopharm...	8
8	Sputnik V	8
9	Moderna, Pfizer/BioNTech	6

# The top vaccines by countries. 
vaccine1 = countryvaccinationbyMan1["vaccine"].value_counts().reset_index()
vaccine1.columns = ['popular Vaccines','Number of Countries']
vaccine1.nlargest(10,'Number of Countries')

	popular Vaccines	Number of Countries
0	Pfizer/BioNTech	10
1	Johnson&Johnson	9
2	Moderna	9
3	Oxford/AstraZeneca	9
4	Sinovac	1

#This plot represents the total number of vaccines count of each kind vaccination kind across the world. 
#From the below graph we can say that Sinopharm/Beijing, Sinopharm/Wuhan, Sinovacn is used to vaccinated more people
# and pfizer /biotech and sputnik V are used less to vaccinate.


plt.figure(figsize=(10,7))
grp = ['country', 'total_vaccinations', 'iso_code', 'vaccines']
vacc_no = countryvaccination[grp].groupby('vaccines').max().sort_values('total_vaccinations', ascending=False).dropna(subset=['total_vaccinations'])


plt.bar(vacc_no.index, vacc_no.total_vaccinations, color ='g')

plt.title('Various categories of COVID-19 vaccines offered')
plt.xticks(rotation = 90)
plt.ylabel('Number of people vaccinated')
plt.xlabel('Vaccines')
plt.show();

#Plotting total number of vaccinations w.r.t manufacturers. From thsi plot we can say that pfizer/biotech and Moderna are used
# more for the vaccination. This dataset only consists of 10 countries for example, USA, Chile, and Romania etc.

x= manuVac.groupby("vaccine")["total_vaccinations"].sum().sort_values(ascending= False)
import seaborn as sns
plt.figure(figsize= (15,10))
ax= sns.barplot(x.index,x.values)
ax.set_ylabel("total vaccinations")
ax.set_xlabel("Manufracturer")  
ax.set_title("Share of Vaccination accoding to manufracturer ") 
plt.show() 

# This plot represents the 10 countries w.r.t total vaccinations.

z= vacc.groupby("country")["total_vaccinations"].max().sort_values(ascending= False)

z['United States']=z['United States']/4
z.to_frame( ) 
countries={'Germany', 'France', 'Italy', 'Chile', 'Romania', 'Czechia', 'Lithuania', 'Latvia' ,'Iceland'}
a = z[z.index.str.contains('United States')]

for x in countries:

    a=a.append(z[z.index.str.contains(x) ])
    
a=a.sort_values(ascending= False)

plt.figure(figsize= (15,10))
ax= sns.barplot(a.index,a.values)
ax.set_ylabel("total vaccinations")
ax.set_xlabel("Country")  
ax.set_title("Country wise vaccination (USA normalized: Multiply by 4 for actual number)") 
# plt.show()

# Creating dataframes to get the total vaccinationd and vaccines according to the manufacturer in each country.
usaVaccName=manuVac[manuVac.location.str.contains('United States')]
gerVaccName=manuVac[manuVac.location.str.contains('Germany')]
itaVaccName=manuVac[manuVac.location.str.contains('Italy')]
litVaccName=manuVac[manuVac.location.str.contains('Lithuania')]
fraVaccName=manuVac[manuVac.location.str.contains('France')]
czeVaccName=manuVac[manuVac.location.str.contains('Czechia')]
latVaccName=manuVac[manuVac.location.str.contains('Latvia')]
chiVaccName=manuVac[manuVac.location.str.contains('Chile')]
romVaccName=manuVac[manuVac.location.str.contains('Romania')]
iceVaccName=manuVac[manuVac.location.str.contains('Iceland')]

gerVaccNames=gerVaccName.groupby("vaccine")["total_vaccinations"].sum().sort_values(ascending= False)
latVaccNames=latVaccName.groupby("vaccine")["total_vaccinations"].sum().sort_values(ascending= False)
usaVaccNames=usaVaccName.groupby("vaccine")["total_vaccinations"].sum().sort_values(ascending= False)
itaVaccNames=itaVaccName.groupby("vaccine")["total_vaccinations"].sum().sort_values(ascending= False)
fraVaccNames=fraVaccName.groupby("vaccine")["total_vaccinations"].sum().sort_values(ascending= False)
czeVaccNames=latVaccName.groupby("vaccine")["total_vaccinations"].sum().sort_values(ascending= False)
chiVaccNames=chiVaccName.groupby("vaccine")["total_vaccinations"].sum().sort_values(ascending= False)
romVaccNames=romVaccName.groupby("vaccine")["total_vaccinations"].sum().sort_values(ascending= False)
iceVaccNames=iceVaccName.groupby("vaccine")["total_vaccinations"].sum().sort_values(ascending= False)
litVaccNames=litVaccName.groupby("vaccine")["total_vaccinations"].sum().sort_values(ascending= False)

#For these 10 countires, we have plotted the Share of each vaccination used. 
#Vaccine rate of each manufacturer in respective countries.
#With respect to those 10 countries,
#we can see the pfizer/biotech is common and used more than 50% to vaccinate in all the respected countries.
# We can see that Chile has used the vaccine manufactured by China.

fig, ((ax1, ax2), (ax3,ax4), (ax5,ax6), (ax7, ax8),(ax9,ax10)) = plt.subplots(nrows=5, ncols=2,figsize=(15,20))
ax1.pie(gerVaccNames, labels = gerVaccNames.index,autopct='%1.2f%%')
ax1.set_title("Germany Vaccination Companies") 
ax2.pie(usaVaccNames, labels = usaVaccNames.index,autopct='%1.2f%%') 
ax2.set_title("USA Vaccination Companies") 
ax3.pie(chiVaccNames, labels = chiVaccNames.index,autopct='%1.2f%%') 
ax3.set_title("Chile Vaccination Companies") 
ax4.pie(itaVaccNames, labels = itaVaccNames.index,autopct='%1.2f%%')
ax4.set_title("Italy Vaccination Companies") 
ax5.pie(fraVaccNames, labels = fraVaccNames.index,autopct='%1.2f%%') 
ax5.set_title("France Vaccination Companies") 
ax6.pie(czeVaccNames, labels = czeVaccNames.index,autopct='%1.2f%%')
ax6.set_title("Czechia Vaccination Companies") 
ax7.pie(latVaccNames, labels = latVaccNames.index,autopct='%1.2f%%') 
ax7.set_title("Latvia Vaccination Companies") 
ax8.pie(romVaccNames, labels = romVaccNames.index,autopct='%1.2f%%') 
ax8.set_title("Romania Vaccination Companies") 
ax9.pie(litVaccNames, labels = litVaccNames.index,autopct='%1.2f%%') 
ax9.set_title("Lithuania Vaccination Companies")
ax10.pie(iceVaccNames, labels = iceVaccNames.index,autopct='%1.2f%%') 
ax10.set_title("Iceland Vaccination Companies")

plt.tight_layout()
plt.show() 

From the above plots we can figure out that which countries are using which vaccinations. Also we see plots w.r.t manufacturer and vaccine brand spread across the world. China use Sinopharm/Beijing, Sinopharm/Wuhan, Sinovacn to vaccinate more people. Johnson & Johnson is the vaccination used to vaccinate during emergency.

Usecase 3: What are the main clusters among the countries and how can they be characterized?

Cluster Kind 1

# Listwise countries represents for each vaccine used.
#From these clusters we can see that Oxford/Astrazeneca vaccine is used by many of the countries and 
#vaccines like abdalo and cassino are used in a particular country.

vaccc = countryvaccination1["vaccines"].unique()
for i in vaccc:
    c = list(countryvaccination1[countryvaccination1["vaccines"] == i]['country'])
    print(f"Vaccine names: {i}n \n Used by countries: {c}\n\n")
    print('<>'*50)

Vaccine names: Oxford/AstraZeneca, Pfizer/BioNTech, Sinopharm/Beijingn 
 Used by countries: ['Afghanistan', 'Maldives', 'Peru']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Oxford/AstraZeneca, Pfizer/BioNTech, Sinovac, Sputnik Vn 
 Used by countries: ['Albania', 'Bosnia and Herzegovina']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Sputnik Vn 
 Used by countries: ['Algeria', 'Armenia', 'Belarus', 'Guinea', 'Kazakhstan', 'Paraguay', 'Syria', 'Venezuela']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Oxford/AstraZeneca, Pfizer/BioNTechn 
 Used by countries: ['Andorra', 'Australia', 'Cayman Islands', 'Costa Rica', 'Isle of Man', 'Oman', 'Panama', 'Saudi Arabia', 'Slovenia', 'South Korea', 'Sweden']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Oxford/AstraZenecan 
 Used by countries: ['Angola', 'Anguilla', 'Antigua and Barbuda', 'Bahamas', 'Bangladesh', 'Barbados', 'Belize', 'Bhutan', 'Botswana', 'Brunei', 'Dominica', 'Eswatini', 'Ethiopia', 'Falkland Islands', 'Fiji', 'French Polynesia', 'Gambia', 'Georgia', 'Ghana', 'Grenada', 'Guyana', 'Jamaica', 'Lesotho', 'Liberia', 'Madagascar', 'Malawi', 'Mali', 'Montserrat', 'Myanmar', 'Nauru', 'New Caledonia', 'Nigeria', 'Papua New Guinea', 'Saint Helena', 'Saint Lucia', 'Samoa', 'Sierra Leone', 'Solomon Islands', 'South Sudan', 'Sudan', 'Suriname', 'Taiwan', 'Tajikistan', 'Togo', 'Tonga', 'Trinidad and Tobago', 'Tuvalu', 'Uganda', 'Uzbekistan', 'Vietnam', 'Yemen', 'Zambia']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Oxford/AstraZeneca, Sinopharm/Beijing, Sputnik Vn 
 Used by countries: ['Argentina', 'Djibouti']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Pfizer/BioNTechn 
 Used by countries: ['Aruba', 'Bermuda', 'Gibraltar', 'Greenland', 'Japan', 'Kuwait', 'Monaco', 'New Zealand', 'Qatar', 'Slovakia']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Johnson&Johnson, Moderna, Oxford/AstraZeneca, Pfizer/BioNTechn 
 Used by countries: ['Austria', 'Belgium', 'Bulgaria', 'France', 'Germany', 'Ireland', 'Italy', 'Latvia', 'Lithuania', 'Netherlands', 'Poland', 'Romania', 'Spain']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Oxford/AstraZeneca, Sinovacn 
 Used by countries: ['Azerbaijan', 'Benin', 'Indonesia', 'Thailand']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Oxford/AstraZeneca, Pfizer/BioNTech, Sinopharm/Beijing, Sputnik Vn 
 Used by countries: ['Bahrain', 'Bolivia', 'Lebanon', 'Moldova', 'Mongolia', 'Montenegro', 'North Macedonia', 'Serbia']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Oxford/AstraZeneca, Pfizer/BioNTech, Sinovacn 
 Used by countries: ['Brazil', 'Chile', 'Colombia', 'Ecuador', 'El Salvador', 'Malaysia', 'Ukraine', 'Uruguay']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Oxford/AstraZeneca, Sinopharm/Beijing, Sinovacn 
 Used by countries: ['Cambodia', 'Dominican Republic', 'Somalia']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Sinopharm/Beijingn 
 Used by countries: ['Cameroon', 'Equatorial Guinea', 'Gabon', 'Mauritania', 'Senegal', 'Zimbabwe']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Moderna, Oxford/AstraZeneca, Pfizer/BioNTechn 
 Used by countries: ['Canada', 'Croatia', 'Estonia', 'Finland', 'Greece', 'Iceland', 'Luxembourg', 'Malta', 'Norway', 'Portugal', 'Rwanda', 'United Kingdom']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Covaxin, Oxford/AstraZenecan 
 Used by countries: ['Central African Republic', 'India']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Sinopharm/Beijing, Sinopharm/Wuhan, Sinovacn 
 Used by countries: ['China']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Covaxin, Oxford/AstraZeneca, Sinopharm/Beijingn 
 Used by countries: ['Comoros', 'Mauritius']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Moderna, Oxford/AstraZeneca, Sinopharm/Beijing, Sputnik Vn 
 Used by countries: ['Congo']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Abdalan 
 Used by countries: ['Cuba']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Johnson&Johnson, Oxford/AstraZeneca, Pfizer/BioNTechn 
 Used by countries: ['Cyprus']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Moderna, Pfizer/BioNTechn 
 Used by countries: ['Denmark', 'Faeroe Islands', 'Israel', 'Liechtenstein', 'Singapore', 'Switzerland']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Oxford/AstraZeneca, Sinopharm/Beijing, Sinovac, Sputnik Vn 
 Used by countries: ['Egypt']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Moderna, Oxford/AstraZenecan 
 Used by countries: ['Guatemala', 'Honduras']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Oxford/AstraZeneca, Sinopharm/Beijingn 
 Used by countries: ['Guinea-Bissau', 'Iraq', 'Morocco', 'Mozambique', 'Namibia', 'Nepal', 'Niger', 'Seychelles', 'Sri Lanka']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Pfizer/BioNTech, Sinovacn 
 Used by countries: ['Hong Kong', 'Turkey']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Moderna, Oxford/AstraZeneca, Pfizer/BioNTech, Sinopharm/Beijing, Sputnik Vn 
 Used by countries: ['Hungary']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Covaxin, Oxford/AstraZeneca, Sinopharm/Beijing, Sputnik Vn 
 Used by countries: ['Iran']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Pfizer/BioNTech, Sinopharm/Beijingn 
 Used by countries: ['Jordan', 'Macao']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Oxford/AstraZeneca, Sputnik Vn 
 Used by countries: ['Kenya', 'Nicaragua']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Sinopharm/Beijing, Sputnik Vn 
 Used by countries: ['Kyrgyzstan', 'Laos']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Johnson&Johnson, Moderna, Oxford/AstraZeneca, Pfizer/BioNTech, Sputnik Vn 
 Used by countries: ['Libya']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: CanSino, Oxford/AstraZeneca, Pfizer/BioNTech, Sinovac, Sputnik Vn 
 Used by countries: ['Mexico']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: CanSino, Sinopharm/Beijing, Sinovac, Sputnik Vn 
 Used by countries: ['Pakistan']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Oxford/AstraZeneca, Sinovac, Sputnik Vn 
 Used by countries: ['Philippines']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: EpiVacCorona, Sputnik Vn 
 Used by countries: ['Russia']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Pfizer/BioNTech, Sputnik Vn 
 Used by countries: ['San Marino']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Johnson&Johnsonn 
 Used by countries: ['South Africa']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Pfizer/BioNTech, Sinovac, Sputnik Vn 
 Used by countries: ['Tunisia']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: EpiVacCorona, Oxford/AstraZeneca, Sinopharm/Beijing, Sputnik Vn 
 Used by countries: ['Turkmenistan']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Oxford/AstraZeneca, Pfizer/BioNTech, Sinopharm/Beijing, Sinopharm/Wuhan, Sputnik Vn 
 Used by countries: ['United Arab Emirates']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Vaccine names: Johnson&Johnson, Moderna, Pfizer/BioNTechn 
 Used by countries: ['United States']


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>

cv = cv.fillna(0)

cv2 = cv.groupby(["country",'iso_code','vaccines'])['total_vaccinations',
                     'people_vaccinated','people_fully_vaccinated',
                    'daily_vaccinations','total_vaccinations_per_hundred',
                    'people_vaccinated_per_hundred',"people_fully_vaccinated_per_hundred"
,'daily_vaccinations_per_million'].max().reset_index()
cv2.head()

	country	iso_code	vaccines	total_vaccinations	people_vaccinated	people_fully_vaccinated	daily_vaccinations	total_vaccinations_per_hundred	people_vaccinated_per_hundred	people_fully_vaccinated_per_hundred	daily_vaccinations_per_million
0	Afghanistan	AFG	Oxford/AstraZeneca, Pfizer/BioNTech, Sinopharm...	504502.0	448878.0	55624.0	13921.0	1.30	1.15	0.14	358.0
1	Albania	ALB	Oxford/AstraZeneca, Pfizer/BioNTech, Sinovac, ...	632676.0	444755.0	187921.0	17565.0	21.98	15.45	6.53	6104.0
2	Algeria	DZA	Sputnik V	75000.0	0.0	0.0	3748.0	0.17	0.00	0.00	85.0
3	Andorra	AND	Oxford/AstraZeneca, Pfizer/BioNTech	31633.0	26931.0	4702.0	1182.0	40.94	34.86	6.09	15298.0
4	Angola	AGO	Oxford/AstraZeneca	626572.0	586377.0	40195.0	18751.0	1.91	1.78	0.12	571.0

# Dataframe for vaccination per 100.
cvcluster = cv2[['people_fully_vaccinated_per_hundred','total_vaccinations_per_hundred']]

#Dataframe display for clustering
cvcluster.head()

	people_fully_vaccinated_per_hundred	total_vaccinations_per_hundred
0	0.14	1.30
1	6.53	21.98
2	0.00	0.17
3	6.09	40.94
4	0.12	1.91

cvcluster=cvcluster.loc[(cvcluster!=0.0).any(axis=1)]

# Normalizing the values
scaler = StandardScaler()
cvcluster1 = scaler.fit_transform(cvcluster)

# Elbow method to find out the number of clusters to choose.
distortions = []
K = range(1,10)
for k in K:
    kmeanModel = KMeans(n_clusters=k)
    kmeanModel.fit(cvcluster1)
    distortions.append(kmeanModel.inertia_)
plt.figure(figsize=(16,8))
plt.plot(K, distortions, 'bx-')
plt.xlabel('k')
plt.ylabel('Distortion')
plt.title('The Elbow Method showing the optimal k')
plt.show()

#Applying K-Means clustering. 
kmeans = KMeans(n_clusters=3)
kmeans.fit(cvcluster1)

KMeans(n_clusters=3)

kmeans.labels_

array([1, 1, 1, 0, 1, 0, 1, 1, 1, 0, 1, 0, 1, 1, 2, 1, 0, 1, 0, 1, 1, 2,
       0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 2, 1, 2, 1, 1, 1, 1, 1, 1, 1,
       1, 0, 0, 0, 0, 1, 0, 1, 1, 1, 1, 0, 1, 0, 1, 1, 0, 2, 1, 0, 0, 0,
       1, 1, 1, 0, 1, 2, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 0, 1, 1, 1, 1,
       0, 0, 2, 0, 1, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0,
       1, 1, 1, 0, 1, 2, 1, 1, 1, 1, 0, 0, 1, 0, 0, 1, 1, 1, 1, 1, 0, 0,
       1, 1, 1, 1, 1, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 1, 1,
       2, 1, 1, 1, 1, 2, 1, 1, 0, 1, 0, 2, 1, 0, 0, 0, 0, 1, 1, 1, 1, 1,
       0, 1, 1, 1, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 1, 1, 1, 2,
       0, 2, 0, 1, 1, 1, 0, 0, 1, 1, 1])

cv2['Clustered'] = pd.Series(kmeans.labels_, index=cvcluster.index)

cv2['Clustered'].value_counts()

1.0    136
0.0     59
2.0     14
Name: Clustered, dtype: int64

cv2_0=cv2[cv2.Clustered == 0]
cv2_1=cv2[cv2.Clustered == 1]
cv2_2=cv2[cv2.Clustered == 2]

Cluster Kind 2

# This tree map represents the cluster 0 , which has the countries with average number of total vaccinations per hundred rate
# and the prople fully vaccinated rate. We can observe that this cluster has the countires from Europe continent and also has the
# moderate population

fig = px.treemap(cv2_0,values = 'total_vaccinations_per_hundred',
                 path = ['country','people_fully_vaccinated_per_hundred'],
                 title="Total Vaccinations per hundred for a country")
fig.show()

# This tree map has the countries which having the less total vaccinations per hundred rate and
#also people fully vaccinated rate. These countries mainly belongs to Africa and Asia continent.

fig = px.treemap(cv2_1,values = 'total_vaccinations_per_hundred',
                 path = ['country','people_fully_vaccinated_per_hundred'],
                 title="Total Vaccinations per hundred for a country")
fig.show()

# This tree map has the countries having highest number of total vaccinations per hundred
# and people fully vaccinated per hundred. But except United States all other countries have less population and 
# all belongs to MiddleEast continent. 

fig = px.treemap(cv2_2,values = 'total_vaccinations_per_hundred',
                 path = ['country','people_fully_vaccinated_per_hundred'],
                 title="Total Vaccinations per hundred for a country")
fig.show()

The above plots we have got after performing clustering, have 2 kinds of clusters. One with the countires used by vaccine manufacturer and the other with the total number of vaccinations per hundred. We can say that from the cluster 1, Oxford/Astrazeneca vaccine is used by many of the countries and vaccines like abdalo and cassino are used in a particular country.

From the Cluster 2, MidddleeEast countries have the more total vaccinations rate per hundred with less population and the African Countries have the less total vaccination per hundred rate.

Usecase 4: How are the vaccination progress and the number of positive cases, deaths etc. linked in the United States?

# Importing the Worldometer Covid cases dataset. 
USA = pd.read_csv("owid-covid-data.csv")

USA = USA[USA.iso_code == 'USA']

# Importing the USA Covid cases dataset used for the previous task.
USAold = pd.read_csv("2021-04-21_covid19_daily.csv")
USAold.head()

	date	states	positive	negative	pending	hospitalizedCurrently	hospitalizedCumulative	inIcuCurrently	inIcuCumulative	onVentilatorCurrently	...	totalTestResults	lastModified	total	posNeg	deathIncrease	hospitalizedIncrease	negativeIncrease	positiveIncrease	totalTestResultsIncrease	hash
0	20201206	56	14534035	161986294	13592.0	101487.0	585676.0	20145.0	31946.0	7094.0	...	204063869	2020-12-06T24:00:00Z	0	0	1138	2256	1172590	176771	1634532	9cf16504f91958e803a2197daf8c2528a4eddc18
1	20201205	56	14357264	160813704	13433.0	101190.0	583420.0	19950.0	31831.0	7005.0	...	202429337	2020-12-05T24:00:00Z	0	0	2445	3316	1526995	211073	2169756	6249216c5f097c94ce33a811dab011a483a42404
2	20201204	56	14146191	159286709	12714.0	101276.0	580104.0	19858.0	31608.0	6999.0	...	200259581	2020-12-04T24:00:00Z	0	0	2563	4652	1260657	224831	1854869	ae30ea088584335ba4d57ee927f8dbda6add74db
3	20201203	56	13921360	158026052	15106.0	100755.0	575452.0	19723.0	31276.0	6867.0	...	198404712	2020-12-03T24:00:00Z	0	0	2706	5331	1238465	210204	1828230	0f253d185ecb336cdd18a4c61996eda1b7eef13b
4	20201202	56	13711156	156787587	14368.0	100322.0	570121.0	19680.0	31038.0	6855.0	...	196576482	2020-12-02T24:00:00Z	0	0	2733	5028	982032	195796	1459202	477c17b6302d0485195e77ecf8270a974f7a3c82

5 rows × 25 columns

USAold = USAold[['date','positiveIncrease','deathIncrease']]

USAold['new_cases'] = USAold[['positiveIncrease']]
USAold['new_deaths'] = USAold[['deathIncrease']]
USAold['date'] = pd.to_datetime(USAold['date'].astype(str), format='%Y/%m/%d')

USAold = USAold.sort_values(by='date')

USAold = USAold[['date','new_cases','new_deaths']] 

start_date='2020-12-07'
end_date = '2021-05-13'
USA['date'] = pd.to_datetime(USA['date'])
USAnew=USA.loc[USA.date >= pd.to_datetime(start_date)]
USAnew=USAnew.loc[USA.date <= pd.to_datetime(end_date)]
USAnew.head()

	iso_code	continent	location	date	total_cases	new_cases	new_cases_smoothed	total_deaths	new_deaths	new_deaths_smoothed	...	gdp_per_capita	extreme_poverty	cardiovasc_death_rate	diabetes_prevalence	female_smokers	male_smokers	handwashing_facilities	hospital_beds_per_thousand	life_expectancy	human_development_index
87830	USA	North America	United States	2020-12-07	15108918.0	194858.0	205512.286	287342.0	1606.0	2331.714	...	54225.446	1.2	151.089	10.79	19.1	24.6	NaN	2.77	78.86	0.926
87831	USA	North America	United States	2020-12-08	15333410.0	224492.0	210694.143	289963.0	2621.0	2342.429	...	54225.446	1.2	151.089	10.79	19.1	24.6	NaN	2.77	78.86	0.926
87832	USA	North America	United States	2020-12-09	15555949.0	222539.0	213548.714	293146.0	3183.0	2395.000	...	54225.446	1.2	151.089	10.79	19.1	24.6	NaN	2.77	78.86	0.926
87833	USA	North America	United States	2020-12-10	15787464.0	231515.0	214677.571	296141.0	2995.0	2400.571	...	54225.446	1.2	151.089	10.79	19.1	24.6	NaN	2.77	78.86	0.926
87834	USA	North America	United States	2020-12-11	16027441.0	239977.0	215705.000	299559.0	3418.0	2503.286	...	54225.446	1.2	151.089	10.79	19.1	24.6	NaN	2.77	78.86	0.926

5 rows × 59 columns

USAnew = USAnew[['date','new_cases','new_deaths']]

#USA new dataset from december 2020 to may 2021
USAnew.head()

	date	new_cases	new_deaths
87830	2020-12-07	194858.0	1606.0
87831	2020-12-08	224492.0	2621.0
87832	2020-12-09	222539.0	3183.0
87833	2020-12-10	231515.0	2995.0
87834	2020-12-11	239977.0	3418.0

frames = [USAold, USAnew]
  
USA1 = pd.concat(frames)

#USA old dataset from Jan 2020 to Dec 2020
USA1.head()

	date	new_cases	new_deaths
319	2020-01-22	0.0	0.0
318	2020-01-23	0.0	0.0
317	2020-01-24	0.0	0.0
316	2020-01-25	0.0	0.0
315	2020-01-26	0.0	0.0

UnitedStatesnew = UnitedStates[['date','daily_vaccinations']]

USA1.reset_index(drop=True, inplace=True)

UnitedStatesnew.reset_index(drop=True, inplace=True)

UnitedStatesnew['date'] = pd.to_datetime(UnitedStatesnew['date'].astype(str), format='%Y/%m/%d')

sub_df = UnitedStatesnew[["date","daily_vaccinations"]]

US = USA1.merge(sub_df, how="outer")

US = US.fillna(0)

#Final USA Dataset after merging old and new ones with the daily vaccinations.
US

	date	new_cases	new_deaths	daily_vaccinations
0	2020-01-22	0.0	0.0	0.0
1	2020-01-23	0.0	0.0	0.0
2	2020-01-24	0.0	0.0	0.0
3	2020-01-25	0.0	0.0	0.0
4	2020-01-26	0.0	0.0	0.0
...	...	...	...	...
473	2021-05-09	21392.0	242.0	2017931.0
474	2021-05-10	36898.0	401.0	2117025.0
475	2021-05-11	33651.0	677.0	2194787.0
476	2021-05-12	35878.0	851.0	2159146.0
477	2021-05-13	38087.0	806.0	2088962.0

478 rows × 4 columns

#This plot represents the Daily Confirmed cases vs Daily Vaccination rate. From this graph we can see that, 
#The daily cases decreases as the vaccination rate increases. 

from plotly.offline import iplot
new_cases = go.Scatter(x=US['date'], y=US['new_cases'], yaxis='y1', name='Daily confirmed cases')
daily_vaccinations = go.Scatter(x=US['date'], y=US['daily_vaccinations'], yaxis='y2', name='Daily vacctinations')

layout_obj = go.Layout(title='USA COVID vs. USA Vaccinations', xaxis=dict(title='Date'), yaxis=dict(title='Daily cases'), yaxis2=dict(title='Vacctinations', side='right', overlaying='y'))
fig = go.Figure(data=[new_cases, daily_vaccinations], layout=layout_obj)
iplot(fig)

#This plot represents the Daily deaths vs Daily Vaccination rate. From this graph we can see that,
#The daily deaths decreases as the vaccination rate increases. 
from plotly.offline import iplot
new_cases = go.Scatter(x=US['date'], y=US['new_deaths'], yaxis='y1', name='Daily deaths')
daily_vaccinations = go.Scatter(x=US['date'], y=US['daily_vaccinations'], yaxis='y2', name='Daily vacctinations')

layout_obj = go.Layout(title='USA COVID vs. USA Vaccinations', xaxis=dict(title='Date'), yaxis=dict(title='Daily deaths'), yaxis2=dict(title='Vacctinations', side='right', overlaying='y'))
fig = go.Figure(data=[new_cases, daily_vaccinations], layout=layout_obj)
iplot(fig)

From the above plots, we can see that the vaccination helps to reduce the number of cases and the deaths. As the Vaccination rate increases, The number of daily positive cases and the deaths decreases.